Using mortality rates as a screening tool: the experience of the Department of Veterans Affairs

Despite widespread agreement that outcome analyses should be followed by in-depth review of the care provided by outlier providers, there are few, if any, reports in the literature describing the effectiveness of this approach to quality improvement. In 1991, the Department of Veterans Affairs (VA) completed a study to measure hospital-specific mortality rates for VA facilities and to use these rates as a screening tool for focused medical record reviews.

The first phase of the project consisted of a comparison of fiscal year 1986 mortality rates for each VA medical center with aggregated VA data, using logistic regression analysis to adjust for patient characteristics related to mortality. In the second phase, medical record reviews of cases from hospitals with elevated levels of mortality were conducted.

The mortality rate analysis identified 44 facilities with rates that were significantly elevated overall or within specific patient categories. Both the mortality rate analysis and the medical record review suggested that quality concerns were more common in predominantly psychiatric medical centers than in other facilities.

The study uncovered important opportunities for improvement, particularly regarding the quality of medical care provided in VA psychiatric hospitals, suggesting there is value in performing intensive review of outlier providers in outcome analyses. However, serious questions remain concerning the validity of the use of administrative data-bases to assess hospitalwide mortality.